Do Images in Emails Trigger Spam Filters? What the Data Actually Shows

Images alone do not trigger spam filters. The relationship between images and image spam filtering is more nuanced than the "never use images in cold email" advice that dominates most guides.

Here is what actually happens: ISPs like Gmail, Outlook, and Yahoo evaluate dozens of signals when deciding whether to deliver an email to the inbox, the promotions tab, or the spam folder. Images are one signal, but they are never the only signal. According to Mailtrap's 2025 Email Deliverability Report, emails with a single small image and a strong sender reputation see virtually no deliverability difference compared to plain text emails from the same sender.

The problem starts when images dominate the email. When the image-to-text ratio tips past 40% image content, ISPs start paying closer attention. When an email is mostly images with very little text (the classic "image spam" pattern used by actual spammers), filters flag it aggressively. This is because spammers historically embedded their message in images to bypass text-based content filters. ISPs learned to treat image-heavy emails with suspicion.

For cold email senders, the decision is straightforward: default to plain text. When you have a legitimate reason to include an image (a product screenshot, a chart, a personalized thumbnail), follow the rules in this guide. Smartlead's SmartDelivery can test your exact email with images against major ISPs before you send to your full list.

How do spam filters evaluate images in emails?

Spam filters analyze six specific image-related signals, and understanding each one tells you exactly what to avoid. The image spam detection technology has evolved significantly since the early days of keyword-based filtering.

• Image-to-text ratio: This is the most important signal. SpamAssassin (used by many corporate mail servers) adds penalty points when the HTML content is predominantly images. The threshold varies, but the general rule is: keep image content below 40% of total email body content. An email that's 80% image and 20% text looks like a promotional flyer, and ISPs treat it as one.
• Image file size: Large images (500KB+) slow load times and trigger filters because legitimate person-to-person emails rarely include large image files. According to Google Postmaster Guidelines, keeping total email size under 102KB (including images) avoids Gmail's clipping behavior where Gmail truncates the message.
• Number of images: Each additional image adds a small amount of filter suspicion. One image is fine. Two images are acceptable. Five or more images in a cold email is a red flag that looks like a newsletter or promotional email, not a personal outreach message.
• Image hosting domain reputation: If your images are hosted on a domain that's been associated with phishing or malware, every email containing those images inherits that negative signal. Always host images on your own domain or a reputable CDN. Never hotlink images from random third-party sites.
• Hidden or single-pixel images (tracking pixels): This is where the deliverability impact gets serious for cold email senders. We cover tracking pixels in detail in the next section.
• Alt text presence: Emails with images that lack alt text attributes score slightly worse in SpamAssassin because legitimate senders typically include alt text for accessibility. Always add descriptive alt text to every image. For more on SpamAssassin scoring, see the SpamAssassin score guide.

Do tracking pixels affect cold email deliverability?

Yes. Tracking pixels have a measurable negative impact on cold email deliverability, and this is the image-related issue that affects the most senders. A tracking pixel is a tiny (usually 1x1 pixel) invisible image embedded in an email to detect when the recipient opens the message.

Here is the problem: every tracking pixel is an external image request. When the recipient opens the email, their email client pings the tracking server to load the pixel. ISPs know what tracking pixels look like. They've cataloged the domains that host them. According to Mailtrap's data, emails with tracking pixels see 8% to 12% lower inbox placement rates in cold outreach compared to identical emails without tracking pixels.

The deliverability impact is worse for cold email than for marketing email because:

• Cold email goes to people who didn't opt in: ISPs apply stricter scrutiny to emails sent to recipients who haven't previously engaged with the sender. A tracking pixel in that context looks like surveillance, not communication.
• Apple Mail Privacy Protection blocks tracking pixels by default: Since iOS 15 (2021), Apple pre-loads all images (including tracking pixels) regardless of whether the user actually opens the email. This means your open rate data from Apple Mail users is unreliable anyway. You're taking a deliverability hit for data that isn't accurate.
• Gmail groups tracked emails differently: Gmail's filtering uses image loading behavior as one signal for categorization. Emails with tracking infrastructure are more likely to land in Promotions than Primary, even when the content is conversational.

What to do about it:

• For cold outreach (first email to a new prospect), consider disabling open tracking entirely. Reply rate is a more reliable success metric anyway.
• If you need open tracking, use it selectively. Enable it for follow-up sequences where you've already established engagement, not for initial cold outreach.
• Smartlead, which functions as an outbound operating system with deliverability at its core, gives you granular control over tracking at the campaign and sequence level, so you can disable tracking on step one and enable it on steps three through five.

The email deliverability guide covers the full picture of signals that affect inbox placement beyond images.

Should you use plain text or HTML for cold email?

Plain text wins for cold email deliverability. The data is clear and consistent across multiple studies. According to G2 user reviews of outbound platforms, teams that switch from HTML templates to plain text emails report 15% to 25% improvements in inbox placement rates.

Why plain text works better for cold outreach:

• It looks like a real email: When you email a colleague, you don't include headers, banners, footer images, and styled buttons. You type text and hit send. Plain text cold emails mimic this pattern, which signals to ISPs that the email is person-to-person communication, not a broadcast.
• No image spam triggers: Plain text emails contain zero images by definition, so none of the image-related spam signals apply. No ratio calculations, no file size concerns, no hosting domain checks.
• Faster rendering: Plain text emails load instantly on every device and email client. HTML emails with images can render differently across Gmail, Outlook, Apple Mail, and mobile clients. Broken rendering leads to immediate deletion.
• Higher reply rates: Beyond deliverability, plain text emails consistently generate higher reply rates in cold outreach. The psychological reason is simple: they look like they were written by a person, for a person. HTML emails look like they were sent by a marketing platform to a list.

When HTML with images makes sense:

• Product demos: If you're showing a specific feature or result (a screenshot, a chart showing ROI), a single embedded image can increase engagement. Keep it under 100KB.
• Follow-up sequences: After the prospect has replied and you've established a conversation, including a product screenshot or case study chart in a later email is reasonable.
• Branded signature images: A small headshot or company logo in your email signature is normal and doesn't trigger image spam filters. Keep signature images under 20KB.

For the full pre-send checklist, see the cold email checklist.

What are the rules for using images in cold email without hurting deliverability?

If you've decided that an image adds genuine value to your email (not just decoration), follow these rules to minimize the deliverability impact. These guidelines are based on SpamAssassin documentation and Google Postmaster Guidelines.

Rule 1: One image maximum per cold email: One image is the threshold where deliverability impact is minimal. Two images are tolerable. Three or more images in a cold email will measurably reduce inbox placement, especially for senders with newer domain reputations.

Rule 2: Keep total email size under 102KB: This is Gmail's clipping threshold. If your email exceeds 102KB (including HTML markup and base64-encoded images), Gmail will truncate it and display a "View entire message" link. Clipped emails have lower engagement rates and the clipping itself is a negative signal.

Rule 3: Maintain a 60/40 text-to-image ratio minimum: At least 60% of your email's visible content should be text. SpamAssassin adds progressively more penalty points as the image ratio increases. The email optimization best practices guide covers content ratio guidelines in detail.

Rule 4: Host images on your own domain: Use your sending domain or a subdomain (images.yourdomain.com) to host email images. This way, image loading contributes to your domain's reputation rather than depending on a third-party CDN's reputation. Never use free image hosting services.

Rule 5: Always include alt text: Every image needs a descriptive alt attribute. Not just for accessibility (though that matters), but because SpamAssassin penalizes images without alt text. Write alt text that describes the image content naturally, not alt text stuffed with keywords.

Rule 6: Compress aggressively: Use WebP or optimized JPEG format. A product screenshot doesn't need to be 2000px wide. Resize to 600px wide maximum and compress to under 50KB per image. The total for all images should stay under 100KB.

Rule 7: Test before sending: Use Smartlead's SmartDelivery to send test versions of your email (with and without images) and compare inbox placement across Gmail, Outlook, and Yahoo. If the image version shows materially lower placement, drop the image. The data should drive the decision, not a guess.

How does sender reputation interact with image filtering?

Sender reputation is the amplifier that makes images either a non-issue or a deliverability killer. The same image in an email from a high-reputation sender lands in the inbox. The identical image in an email from a low-reputation sender lands in spam. ISPs don't evaluate images in isolation; they evaluate images in the context of sender trust.

This is why the "do images cause spam" question doesn't have a simple yes/no answer. It depends on who is sending.

According to Google Postmaster data, senders with "High" reputation scores can include images, HTML formatting, and even tracking pixels with minimal deliverability impact. Senders with "Low" or "Medium" reputation scores see significant drops in inbox placement when they add any of these elements.

For cold email senders, this creates a specific challenge:

• New domains start with no reputation: If you're sending from a domain that's three weeks old and just finished warmup, your reputation is "unknown" at best. Adding images to emails sent from unknown-reputation domains is risky because ISPs have no trust context to offset the image signals.
• Reputation builds through engagement: ISPs track how recipients interact with your emails. Opens, replies, and "not spam" reports build reputation. Deletes, spam reports, and ignores damage it. The warmup guide covers how Smartlead's unlimited warmup builds this engagement pattern before you start sending campaigns.
• Reputation is domain-specific and IP-specific: Smartlead's SmartServers give you dedicated sending infrastructure, which means your reputation isn't affected by other senders sharing the same IP. On shared infrastructure, another sender's poor practices (including heavy image usage) can drag down everyone's deliverability.

The practical takeaway: Build your sender reputation with plain text emails first. Once you have a "High" reputation score (check Google Postmaster Tools), you have more flexibility to include images without deliverability consequences. Starting with images on a fresh domain is asking for problems.

What about GIFs, videos, and rich media in cold email?

Every rich media format triggers more image spam signals than a static image, so the rules are stricter.

• GIFs: Animated GIFs increase email size dramatically (often 500KB to 2MB), blow past Gmail's clipping threshold, and trigger aggressive spam filtering. Some email clients don't render GIFs at all, showing only the first frame or a broken image icon. Use GIFs only in follow-up emails to engaged prospects, never in cold outreach.
• Embedded video: You cannot embed actual video in email. The HTML5 video tag is not supported by any major email client except Apple Mail. What most tools do is embed a static image that links to a video. That's just an image with a link, and the regular image rules apply. Keep the thumbnail under 50KB and test placement with SmartDelivery.
• Interactive elements: CSS animations, hover effects, and interactive carousels are supported by very few email clients and are treated as suspicious HTML by most spam filters. Avoid them entirely for cold email.
• Personalized images (Lemlist-style): Services that generate personalized images (prospect's name on a whiteboard, company logo on a mockup) add one external image per email. The deliverability impact is moderate because the image is hosted on the personalization service's domain, which is usually well-maintained. But you're still adding an image to a cold email, which means the sender reputation interaction applies. Test before scaling.

Mailtrap's data shows that emails with any rich media element (GIFs, video thumbnails, interactive elements) see 18% lower inbox placement than text-only equivalents in cold outreach contexts. For landing in the primary inbox, see the 7 ways to land in the primary inbox guide.

How do you test whether images are hurting your specific campaigns?

Don't guess. Test. Smartlead gives you the tools to answer this question definitively for your specific campaigns, your specific audience, and your specific sender reputation.

• Step 1: Run an A/B test: Create two variants of the same email. Variant A is plain text with no images. Variant B includes the image you want to use. Send both variants to the same audience segment with the same subject line. Smartlead's A/B testing distributes variants evenly and tracks performance differences.
• Step 2: Check inbox placement before sending: Use SmartDelivery to test both variants against Gmail, Outlook, and Yahoo. This shows you where each version lands before it reaches your actual prospects. If Variant B (with image) shows 90%+ inbox placement, the image is safe. If it drops below 80%, reconsider.
• Step 3: Monitor over time: Deliverability changes as ISPs update their algorithms and as your sender reputation evolves. An image that was fine last month might cause issues this month if your reputation dipped. Check monthly.
• Step 4: Segment by ISP: Your images might be fine for Gmail recipients but problematic for Outlook recipients (corporate mail servers running SpamAssassin tend to be stricter about image content). Smartlead lets you segment campaign data by recipient ISP so you can make ISP-specific decisions.

The teams that consistently achieve the best deliverability are the ones that test everything and let data drive decisions. The Smartlead Master Inbox centralizes all reply data so you can compare reply rates between image and no-image variants across your entire campaign portfolio.

Frequently Asked Questions

1. Do images automatically send emails to spam?

No. Images do not automatically trigger spam filters. What triggers filters is the ratio of image content to text content, the total size of images, the number of images, and the reputation of the sender and image hosting domain. A single optimized image (under 50KB, with alt text, hosted on your own domain) has minimal impact on deliverability for senders with established reputations. The image spam signal is a composite of multiple factors, not a binary image-yes/image-no decision.

2. Should I disable open tracking for cold email?

For initial cold outreach (first email to a new prospect), disabling open tracking improves deliverability by removing the tracking pixel. Mailtrap's data shows 8% to 12% better inbox placement without tracking pixels. Reply rate is a more reliable metric than open rate anyway, especially since Apple Mail Privacy Protection inflates open rate data. Smartlead lets you control tracking at the sequence step level, so you can disable tracking on step one and enable it on later follow-ups.

3. Is plain text always better than HTML for cold email?

For cold outreach to new prospects, plain text outperforms HTML on deliverability and reply rates in almost every test. The exceptions are follow-up emails to already-engaged prospects (where a product screenshot adds value) and emails where a visual element is central to the pitch (personalized video thumbnails, custom charts). Even in those cases, test with Smartlead's SmartDelivery before scaling.

4. How many images can I include in a cold email?

One image maximum for cold outreach. Two images if absolutely necessary. More than two images makes your email look like a newsletter or promotional blast, which triggers image spam filters and signals to ISPs that this is marketing email, not person-to-person communication. Keep total image payload under 100KB and maintain at least 60% text content.

5. Do email signature images affect deliverability?

Small signature images (headshot, company logo) under 20KB have negligible deliverability impact for senders with established reputations. For new domains still building reputation, even signature images add unnecessary risk. Start with a text-only signature during the warmup and early campaign phases, then add a small image once your reputation is established.

6. How do I check if images are causing my emails to land in spam?

Use Smartlead's SmartDelivery to test identical emails with and without images across Gmail, Outlook, and Yahoo. Compare inbox placement rates. If the image version shows more than a 5% drop in placement, the image is hurting you. Also check your SpamAssassin score for both versions. SpamAssassin adds specific penalty points for image-to-text ratio violations.